CHAFF CLOUD SIGNATURE

Claud C. Pinson
Pinson Associates, Inc.
Austin
Texas

Summarv

Chaff deployed from aircraft forms a

cloud that grows very rapidly. Typically, it
expands by a factor of 2 x 10.^ or more during
the first .25 seconds after deployment. The
turbulent flow around and behind the aircraft
causes the chaff cloud to grow rapidly and to
have RCS values that vary rapidly with time.
Changes in RCS of 30 dB or more in less than a
millisecond are common, and changes of over 40
dB have been observed. After the chaff cloud
is out of the wake of the aircraft, it con­
tinues to grow but at a much slower rate. The
direction of growth is primarily with the wind
and in the vertical plane. The RCS variations
with time are reduced, however, they are still
significant.

Measurements made over the past two

years using the Real Time Radar Cross Section
(RCS) Processor have provided some insight
into the chaff cloud, growth and true radar
signature of the cloud. This instrument, with
an 80 dB dynamic range, 15 sampling gates and
pulse-to-pulse RCS measurement capability pro­
vides the data necessary to study the statis­
tics of. the chaff signature. It is known that
the wide variations in RCS that occur in less
than a millisecond is due to the relative
phase changes between the signals from the in­
dividual elements in the cloud. It is b e ­
lieved that the resultant signal, as seen by
the radar, has large random variations in both
amplitude and phase which results in a complex
modulation on the chaff signal. The varia­
tions in amplitude have been proven, and work
is untier w a y to define the characteristics of
the phase changes of the chaff signal.

For this paper, we are indebted to the

Chaff Countermeasures Section of the EW divi­
sion of the Naval Research Laboratory. They
sponsored the design, development and fabri­
cation of the Real Time Radar Cross.Section
(RCS) Processor and permitted us to use some
of the data from field measurements to pre­
pare this paper. Mr. Victor J. Kutsch has
been our major sponsor and he has been sup­
ported by Mr. Richard L. Bloom.

Let us start with a review of a few FIGURE 1

fundamentals concerning chaff. Figure 1 is a CHAFF UNIT
typical chaff unit. It was designed to'be
used against radars operating in the E, F, G,
and I-bands. It has about 750,000 half-wave travel. During the first quarter of a second
dipoles for the E and F-band, approximately after deployment, the dipoles separate and
750,000 half-wave dipoles for the G-band and 3
occupy a volume of 700 to 1500 ft , and their
approximately 2,250,000 half-wave dipoles for velocities are distributed in some weird and
the I-band. The unit is produced for use as unknown fashion between essentially zero and
an aircraft penetration aid. When it is ejec­ the maximum airflow in the area of the cloud.
ted from the dispenser, it has a volume of Thus, the volume of the chaff expands by,a
approximately 6.7 in. and the dipoles have a factor of 1.8 x 10^ to 4 x 10^ in the first
velocity on the order of 80 to 90 feet per .25 seconds. For the next few seconds, the
second in the direction of ejection, and the cloud continues, to grow under the influence of
velocity of the aircraft (nominally 800 feet the turbulent airflow generated by the aircraft.
per second) in the direction of the aircraft As time progresses, the volume of the cloud

2 A I b 1

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.
 continues to grow due to the natural air describe the true signature of the chaff cloud
movements and gravity, with lesser effects as seen by the radar.
due to the turbulence. In this latter life
cycle (one minute to one hour, or longer), the The true signature of the chaff cloud is
growth is primarily in the direction of the obviously the vectorial sum of all of the re­
wind and in the vertical plane. Some five turn signals from all of the individual di­
minutes after deployment, the dipoles may be poles in the cloud. Assuming no changes in
distributed over a volume of approximately the radar transmitted power, the amplitude of
6 x 107 ft3 . the signal from the individual dipoles will
change with time depending upon the orienta­
When chaff is used as an aircraft tion with respect to the radar wave and the
penetration aid, there are two time periods in range separating the dipoles from the radar.
which we are primarily interested in the sig­ Since we are interested in time intervals of a
nature of the chaff cloud. The first time few seconds at most, for practical purposes,
period is from ejection to approximately two the range variations can be ignored and the
seconds after ejection. This is the critical change in the amplitude of the signal from the
time when the chaff is used as a self-protec­ individual dipoles can be related to orienta­
tion countermeasure device and the signature tion. Wi t h respect to the radar signal, the
of the chaff cloud is strongly influenced by phase of the signal from the individual di­
the turbulent airflow around and behind the poles change with, time depending upon their
aircraft. The second time period is one to orientation with respect to the radar waye
sixty minutes after deployment and it is im­ and the range separating the dipoles and the
portant when using chaff as a saturation radar. The change in the range is by far the
countermeasure. In this case, the chaff sig- , major factor in causing rapid phase changes
nature is predominently influenced by the nat­ since only a half wavelength change gives a
ural aerological conditions. It should be re­ 360° change in phase. For reasonable chaff
called that the effectiveness of chaff as a clouds, i.e., number of dipoles, physical area
countermeasure against any radar will depend and airflow conditions, it is impractical, and
upon having the cloud at the right place, at maybe even impossible to calculate the posi­
the right time, and providing the right sig­ tions of the millions of dipoles and to vec­
nature. We choose to define the chaff sig­ tor ially add the signals from each. Thus,
nature in terms of the average RCS of the empirical data must be used to define the chaff
cloud and the modulation characteristics of cloud signatures and how they vary with time.
the signal due to the variation with time of
the amplitude and phase of the signal. About three years ago, Pinson Associates,
Inc., designed and fabricated our first Real
; The backscatter characteristics of Time RCS Processor for the Naval Research Lab­
individual chaff dipoles are well defined by oratory. This equipment could be used to make
antenna theory. Both theoretical calculations pulse-to-pulse RCS measurements at 15 places
and laboratory measurements have been made in range and printout and record the data in
where backscatter or RCS values were deter­ real t i m e . One of the features of the in­
mined as a function of frequency, orientation, strument is an 80 dB dynamic range. Thus,
materials, cross-sectional area and shape, etc. the signal could be measured without any clip­
The theory and laboratory results are in ping. From measurements made with these in­
agreement and are well documented in the lit­ struments, we have found that the pulse-to-
erature and will not, therefore, be discussed pulse amplitude variations of the chaff signal
in this paper. The point of interest is the are far greater than expected. We have also
collective behavior of the millions or more found that in the zero to two second period of
dipoles that form the chaff cloud. the chaff cloud growth, these pulse-to-pulse
amplitude variations are far greater than
It has been traditional to examine the those occurring some five to ten minutes after
collective behavior of the dipoles in a chaff the chaff has been deployed. For the chaff
cloud in the quiet state and to make certain clouds formed immediately behind the aircraft,
assumptions in order to define the signature a change of 30 dB in amplitude in less than
of the chaff signal. Work in this area has one millisecond is common and greater than 10
been done by Tai, Van Vleck, Brown, Richmond dB changes occur a larger percentage of the
and others. 'The assumptions usually made are: time.

a . The dipoles are randomly oriented Before looking at some of the chaff
with respect to the radar wavefront. cloud signature data, let us first consider
the logic of why and how such large variations
b. The dipoles are separated in RCS can occur in less than a millisecond.
sufficiently such that electromagnetic coup­ Consider a chaff cloud that is growing in
ling between the elements can be neglected. cross-sectional area at a half angle of 2° to
5° in the wake of an aircraft where the cross
c. Each of the dipoles is exposed to velocity wind components are up to 90 to 100
the full strength of the radar wave.
feet per second. For the 2° half angle
d. The dipoles are perfect conductors. growth rate and a 600-foot per second aircraft
velocity, the extremities of the cloud are
Using the assumptions listed above, the separating at 1277 centimeters per second,
RCS of the chaff cloud consisting of n reso­ yielding a maximum average phase change be ­
nant ,half-wave dipoles is .15 n \^ where n is tween dipoles of 306° per millisecond for I-
the number of dipoles and \ is the wavelength. band. For the 5° half angle growth rates and
While this expression is very helpful in d e ­ 800 feet per second aircraft velocity, the
signing chaff units, it does not, in any way. maximum average values are 4267 centimeters

2 A I b 2

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.
 per second and 1024 per millisecond, (without clipping) of the chaff cloud 200 feet
respectively. Now, if one only considers the and 300 feet behind a high performance jet air­
flow rates out from vortices or exhaust, the craft have been used to calculate the spectrums
relative velocity of the dipoles with respect of the chaff return signals. In this case,,
to the direction of the radar, could vary the PRF of the radar was 1260 Hz and, there­
from, a -100.feet per second to a positive 100 fore, the spectrum could be plotted' out to 630
feet per second and assume all values in.be­ Hz. The spectrum results are shown in Figure
tween these limits. This corresponds to a 3.
maximum difference.of 6096 centimeters per
second or 1463° per millisecond for three
centimeter or I-band frequencies.

From the discussion above, it is clear

that the relative phase between the signals CM
from, the individual dipoles could vary from 2
0 to well over 3.60° within less than a milli­
second, which offers a potential of a very
large variation in the vectorial, sum of all o
ac
the dipole signals. This indeed is the case
as we will be shown later on. This also
raises an interesting point on the change in
the phase of the composite signal (the vec­
torial sum of all the dipole signals). There
are, at present, no empirical data to show FREQUENCY (Hz)
how the phase of the chaff signal will change
on a pulse-to-pulse basis. It seems logical GATE
that the phase change would be random in mag­
n it u d e and direction, however, the magnitude
and distribution are not readily predictable.
1 see no logical reason why the amplitude
variation and phase variations should be re­
lated in.any ordered fashion. Plans are under
w ay to provide the necessary instrumentation
and to, conduct tests to investigate the pulse-
to-pulse phase changes in chaff cloud returns.
This should pave the way for defining the
actual signature of the chaff cloud.

Many of you may recall seeing the

spectrum of a typical chaff cloud. The typi­
cal one shown in the literature is repeated in
Figure 2.
FREQUENCY (Hz)
GATE 6
FIGURE 3
CHAFF SIGNAL SPECTRUM

Note that the peak values at the higher

frequencies are only down about 6 dB from the
> t
zero value, while the minimum values are down
fez by about 12 d B . The measurements were made in
a dynamic environment where the chaff cloud
VC f : was passing the 100-foot radar range gate at
about 600 feet per second. It is noted that
only approximately 200 pulses are available at
each point for use in defining the spectrum
and, therefore, the values at any given fre­
20 40 60 80 100 quency would change with time and from unit-
CPS to-unit. The figure does indicate the amount
of energy that is contained in the higher fre­
FIGURE 2 quencies.

FREQUENCY SPECTRUMS OF CHAFF Perhaps a mbre interesting way to look

at the pulse-to-pulse variations in RCS of the
Note that the energy is contained in the chaff cloud is to consider some of the statis­
0 to 80 Hz region.- The four curves' are for tics of some individual chaff clouds. In this
four wind conditions ranging from calm to very connection, we chose five individual chaff
gusty conditions. These curves are the re­ clouds for evaluation. All of these units have
sults of measurements made with low PRF radars the same volume of dipoles and all use
and instrumentation having a limited dynamic aluminum-coated glass fibers as the chaff ma ­
range. Furthermore, the chaff dipoles had terial. Two of these chaff clouds have di­
drastically different mechanical and aero­ poles of only one length and, therefore, the
dynamic properties than those in use today. RCS values were measured only at or near the
Also, the chaff cloud has been deployed for resonant frequency. Two of these chaff clouds
several minutes. The pulse-to-pulse RCS data had dipoles cut for three frequencies and the
2 A I b 3

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.
 RCS values measured consisted of at,.or near experiments with higher radar P R F 1s are
resonance response and harmonic response. required to determine the maximum rate of
Measurements on all four of these, units were change in the chaff signal that may result.
made at approximately 200, 900 and 1300 feet
back of the radar centroid of the aircraft The response of an old chaff cloud as
deploying the chaff... This corresponds to shown.by unit 5 in Table I is not so violent.
approximately. .4, 1.6 and 2.2 seconds after It is noted that in any One radar resolution
deployment. The fifth chaff cloud had di­ cell, only approximately 1% of the pulses
poles of only one length (same type of unit have a change in amplitude of 1.0 dB or more,
as the first two listed a bo v e ); however, it and none had changes of 20 dB or more. The
had been deployed for approximately ten min­ PRF of the radar was sufficient to define the
utes, and had grown to occupy four radar re­ maximum rate of change in the signal level
solution cells in range. All of the data from the old chaff cloud. The importance of
were taken at I-band and the PRF of the radar vectorially adding the response .from the,
was 1260 Hz. elements to get the RCS of the composite sig­
nal is clearly shown by the data for unit 5.
The pulse-to-pulse RCS data were First, by comparing the data for gates 1, 2, 3
examined to ensure that the changes in ampli­ and 4, it is clear that each gate had a dif­
tude were due to the signature changes only ferent number of dipoles being illuminated;
and not due to illuminating a different number however, the statistics were essentially the
of dipoles, i.e., all of the dipoles are in same for each. This shows that the statistics
the radar resolution cell during the time of of the signal from the total cloud should be
measurement. For the early chaff cloud condi­ about the. same as for the four parts shown.
tions, this gave four chaff units and three Since the phase relationships of the RCS from
gates or ranges for a total of 12 data each of the four radar resolution cells were
samples. The number of pulses used in each not known, the numerical sum was calculated
case varied from 119 to 216, with 7 of the 12 for each pulse and the statistics of these
varying from 156 to 178. In each case, the numbers are presented on the line marked sum.
maximum, minimum and average RCS values and Note that the maximum change in RCS was only a
the standard deviations were recorded. Also, factor of 2.6 and only .1% had a change of 3
the number of pulses used for each of the 12 dB or more while the individual RCS showed
conditions; the number of pulses where the approximately 14% of the pulses had a change
2 in RCS of 3 dB or more. This indicates the
RCS was less than 1 m ; th e number of con­
deficiency that can result from using short
secutive repeats where the RCS were less than
pulse radars to study chaff, cloud signatures
1 m 3 ; the maximum change in the RCS between unless phase and amplitude measurements are
two consecutive pulses; and the number of con­ made on a pulse-to-pulse basis.
secutive repeated values that were greater
2 The randomness of the amplitude of the
than 1 m were recorded. In addition, the
ratio of standard deviation to the average chaff cloud signal is clearly shown b y the
RCS value; the maximum RCS divided by the data in Table I. There are too few samples
average value; the percentage of time the RCS (two of each type of unit for each time after
changed greater than 3 d B ; the percentage of deployment) to establish any trends. By com­
time the RCS changed greater than 10 dB; the bining the data in Table I with similar data
percentage of time the RCS changed greater from other chaff units measured in the same
than 20 d B ; and the percentage of time the RCS manner, one can draw some general conclusions.
changed less than 2 dB were calculated and re­ Some of these are:
corded. These data are shown in Table I for
all five chaff units. Units 1 and 2 have a . The average RCS values seem to
single frequency dipoles, units 3 and 4 have level out whe n 20 or more units are measured.
multiple frequency dipoles and unit 5 is the
old chaff with single frequency dipoles. For b. Several factors listed in Table I
the fifth unit, data are presented for each of seem to follow the same trends of increasing
the four radar resolution cells and the numeri­ for a short period of time, (.3 to ,6 seconds)
cal sum of the-four radar cross sections com­ after deployment, remain about the same for
puted on a pulse-to-pulse basis. the next two seconds, and then to decrease.
These factors include the standard deviation
The early growth of the chaff cloud as a divided by the average RCS; maximum RCS divi­
function of time is clearly shown for all four ded by the average RCS; maximum change; and
units (1 through 4) by comparing the average the average change from pulse-to-pulse. It is
RCS at the three different time intervals. noted that all of these are related and it is
There are, however, no obvious trends in the believed that the increase takes place con­
pulse-to-pulse RCS variations that occur. It current with the physical cloud growth 'as long
is clear that the pulse-to-pulse RCS varia­ as the cloud is super-saturated with dipoles.
tions are large with approximately 8% of the It then remains unchanged 5 s long as the cloud
consecutive pulses having 20 dB or more varia­ is in the turbulent flow of the. aircraft and
tions and approximately 30% having 10 dB or is not super-saturated. Following the re­
more variations. Approximately 68% of the duction in air turbulence, these factors de ­
consecutive pulses have RCS variations of 3 dB crease in value.
or more, while only 23% of the pulses have a
change in RCS of less than 2 dB. A detailed c. The maximum RCS of any one pulse
study of the changes in the RCS values from seems to increase with increase in average RCS
pulse-to-pulse of the chaff signal from the 12 values. These two increases have not been
data- samples shows that the PRF of the radar tied together in any ordered fopm, i.e., ,
(1260 Hz) was too low to accurately define the linear, constant multiplier, etc..
spectrum of the signal due solely to the
changes im amplitude of the signal. Further d. The number of RCS values less than
2 A I b 4

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.
 TABLE I
RCS STATISTICS

RCS : No . No.
Std Chanqe in RCS
No. Repeats Repeats
Chaff Std No. Max Dev Max % % % %
2 •
, 2 2
Unit Max A vq Dev <1 m Pulses <1 m Chanqe >1 m Avq Avq >3 dB >10 dB >20 dB <2 dB
Unit 1 G-4 147 18.6 27 .7 52 17 7 11 262 .2 2 1.48 7.90 73 .4 37 6 18.1
Unit 1 G-ll 724 52 .1 95.5 35 173 6 1,608.9 3 1.83 13 .89 76.3 38 12 16.2
Unit 1 G-15 836 75.4 139.3 16 119 2 968.9 2 1.84 11.08 66 .4 29 6 22 .7

Unit 2 G-4 89 14.6 16.9 21 211 1 1,100.0 7 1.16 6.10 67 .8 26 12 . 21.3

Unit 2 G-ll 436 57 .5 75.8 10 178 1 1,033 .3 6 1.32 7.58 60 .7 27 5 25.3
Unit 2 G-15 542 60.9 92 .1 18 178 6 1,366.7 11 1.51 8.90 52 .2 16 2 34.8

Unit 3 G-4 197 22 .3 33.6 41 124 . 12 3 50.0 3 1.51 8.83 74 .2 49 7 19.4

Unit 3 G-ll 406 30.8 58.5 48 159 18 673 .8 2 1.90 13 .18 69 .2 31 11 22.0

2 A Ib 5
Unit 3 G-15 542 56.0 87.2 22 156 4 504.8 4 1.56 9 .68 74 .4 26 7 18.6

Unit 4 G-4 89 13 .8 16.7 29 216 6 257.89 10 1.21 6.45 63 .4 24 2 25.0

Unit 4 G-ll 504 52 .9 83.7 20 135 5 1,315.0 2 1.58 9.53 65.2 27 7 25.9
Unit 4 G-15 351 57 .3 76.9 27 158 5 1,315.0 2 1.34 6.13 67 .7 33 16 22 .8

Unit 5 G-l 469 53.3 62.9 24 1032 5 49.4 96 1.18 8.8 19.4 2.7 0 71.0
Unit 5 G-2 724 134 .5 123.7 0 1032 0 620 132 .92 5 .4 10.3 .9 0 80.6
Unit 5 G-3 1039 130.7 174.2 5 1032 0 8.6 141 1.33 8.0 7.2 0 0 82 .5
Unit 5 G-4 158 22 .1 27.0 109 1032 65 19.8 108 1.22 7.1 17.5 .3 0 70.0

Sum 1263 340.81 226.77 0 1032 0 2.6 0 .67 3 .71 0 0 0 96

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.
 2 dispensed from jet aircraft, these amplitude
1 m decreased with any increase in average variations generate changes in RCS of 30 dB or
RCS, however, this decrease was very low as more that are common with 20 d£ or more occur­
long as the chaff was in the turbulent flow of ring approximately 8% of the time, and 10 dB or
the aircraft. A decrease by a factor of about more occurring about 30% of the time. These
30% corresponds to an increase in average RCS variations in RCS generates a spectrum that is
of about 400%. very, wide compared to that conventionally re­
ported in the literature. Further work is re­
e. The distribution of the changes in quired to clearly define the limits of the RCS
RCS between the pulses did not show as much
variations and to define the phase variations
change with cloud growth as the other para­ that occur on a pulse-to-pulse basis.
meters discussed in b above, i.e., the percent
greater than 3 dB, 10 dB and 20 d B .

f. There were no significant

Change
differences in the RCS characteristics from

7 .52

30.49
38.79
29.66

37.44
6.32
41.61
38.22
units having single cuts of dipoles and those

Avg
having multiple cuts of dipoles.

Time and space does not permit a

detailed discussion of the trends that have TJ

RCS
been observed, or the logic that seems to ex­ o

7
8
8

2
2

8
7
13
plain each of them. It is, however, important <N
to note that the change in RCS from one pulse A

in
to the other generates the spectrum of the sig­

> 10 dB
nal due to amplitude variations. The maximum,

Change
minimum and general distribution of these

6
6
32
31

36
28
%

24
26
changes are shown in Table I. The average of
the pulse-to-pulse changes in RCS varied from
13.43 to 81.96 for the 12 samples^ (four units
and three g a t es ) . The overall average change CQ

67 .9
64 .6
69.0
was b y a factor of 36.58 or 15.6 d B .

35
33

77
69
60
It is perhaps important to make a few A
other points about the pulse-to-pulse RCS

6.85
11.80

3.98
14.93

5.60
4.86

10.43
13.59
variations of the chaff cloud, particularly

Max
A vg
when the clouds are in the turbulent flow b e ­
hind the aircraft. These include: the com­
posite data from all four units shown in Table

< 1 m 2 Pulses < 1 m . Change

' 825
325

900
1,100

! 1,125
375
Max

1,367
1,609
I; the aircraft signal alone; the aircraft sig­
nal plus chaff; and the chaff unit that is
only a fraction of the physical size of the

j
ones used in Table I. These data are shown in
Repeats
2

Table II. The first three lines of Table II

Ho.

7
show the composite data for all of the first

6
30
30

17
four units shown in Table I; Most of the gen­
eral trends discussed above can be seen from

!
1
these data. The last three lines in Table II
No.

shows the data for a single unit that has app­

728

194
645
611

291
672

209
202
roximately 12% of the volume of the units
shown in Table I. Comparing these data to
those in Table I or the first three lines of
No.

9
18
83

30
24
143

40
113

Table II, shows that the maximum, minimum and

distribution of the changes in the pulse-to-
pulse RCS values are essentially the same for
the smaller unit as for the normal size unit.
23.5

7.1
13.1
8.9

21.4
80.1

9.3
98.6

The actual range of RCS values are, however,

Dev
Std

quite different. Also, the ratio of.the m a x­

RCS

imum to average RCS decreases as the cloud

becomes sparsely populated, i.e., gates 11 and
56 10.0
100 14.6
724 48.5

52 10.7

4.6
197 16.7

45 11.3
836 61.5

Cn
15 of the small unit. >
<
For comparison, the characteristics of X
48

the aircraft alone, and aircraft plus chaff

ro
2
are shown in Table II. The lower scintillation
+ Chaff

rates and spectrum characteristics of the air­

4 G — 11
Unit

4 G-15

craft as compared to chaff is obvious. It is

4 G-4

Only

noted that the change in the pulse-to-pulse

15
11
4

RCS values of the aircraft and chaff are some

Chaff

greater than the aircraft alone and both lie

Gate
Gate
Gate

between those for the .chaff in the turbulence

All
All
All

A/C
A/C

of the aircraft and the chaff that is out of

the turbulence of the aircraft.

In conclusion, it can be stated that the

pulse-to-pulse amplitude variations of the
chaff cloud depends upon the turbulence in and
around the cloud. For self-protection chaff
2 A I b 6

Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on September 21,2023 at 23:15:44 UTC from IEEE Xplore. Restrictions apply.